A scalable data-distributed algorithm for volume rendering on parallel virtual machines

摘要

An algorithm is presented for volume rendering in environments of parallel virtual machines. To reduce the communication cost, as well as to guarantee the locality of all subtasks, the volume data is divided into and organized as a series of slices. The task subdivision algorithm produces acceptable load balancing by maintaining and employing a database of performance indices. An asynchronous binary method merges all partial images in O (logn) time. An efficient development platform based on TCP/IP and Socket standards can help parallelize various rendering algorithms on virtual machines. Our algorithm was implemented on this platform using the classical client/server paradigm. The scalabilities lities of both task size and host number were tested experimentally.